With the recent explosion of large language models (LLMs), such as Generative Pretrained Transformers (GPT), the need to understand the ability of humans and machines to comprehend semantic language meaning has entered a new phase. This requires interdisciplinary research that bridges the fields of cognitive science and natural language processing (NLP). This pilot study aims to provide insights into individuals' neural states during a semantic relation reading-comprehension task. We propose jointly analyzing LLMs, eye-gaze, and electroencephalographic (EEG) data to study how the brain processes words with varying degrees of relevance to a keyword during reading. We also use a feature engineering approach to improve the fixation-related EEG data classification while participants read words with high versus low relevance to the keyword. The best validation accuracy in this word-level classification is over 60\% across 12 subjects. Words of high relevance to the inference keyword had significantly more eye fixations per word: 1.0584 compared to 0.6576 when excluding no-fixation words, and 1.5126 compared to 1.4026 when including them. This study represents the first attempt to classify brain states at a word level using LLM knowledge. It provides valuable insights into human cognitive abilities and the realm of Artificial General Intelligence (AGI), and offers guidance for developing potential reading-assisted technologies.
Recent years have witnessed increasing interest in few-shot knowledge graph completion (FKGC), which aims to infer unseen query triples for a few-shot relation using a handful of reference triples of the relation. The primary focus of existing FKGC methods lies in learning the relation representations that can reflect the common information shared by the query and reference triples. To this end, these methods learn the embeddings of entities with their direct neighbors, and use the concatenation of the entity embeddings as the relation representations. However, the entity embeddings learned only from direct neighborhoods may have low expressiveness when the entity has sparse neighbors or shares a common local neighborhood with other entities. Moreover, the embeddings of two entities are insufficient to represent the semantic information of their relationship, especially when they have multiple relations. To address these issues, we propose a Relation-Specific Context Learning (RSCL) framework, which exploits graph contexts of triples to capture the semantic information of relations and entities simultaneously. Specifically, we first extract graph contexts for each triple, which can provide long-term entity-relation dependencies. To model the graph contexts, we then develop a hierarchical relation-specific learner to learn global and local relation-specific representations for relations by capturing contextualized information of triples and incorporating local information of entities. Finally, we utilize the learned representations to predict the likelihood of the query triples. Experimental results on two public datasets demonstrate that RSCL outperforms state-of-the-art FKGC methods.
Completion through the embedding representation of the knowledge graph (KGE) has been a research hotspot in recent years. Realistic knowledge graphs are mostly related to time, while most of the existing KGE algorithms ignore the time information. A few existing methods directly or indirectly encode the time information, ignoring the balance of timestamp distribution, which greatly limits the performance of temporal knowledge graph completion (KGC). In this paper, a temporal KGC method is proposed based on the direct encoding time information framework, and a given time slice is treated as the finest granularity for balanced timestamp distribution. A large number of experiments on temporal knowledge graph datasets extracted from the real world demonstrate the effectiveness of our method.
Domain adaptation which pays attention to exploiting the knowledge in source domain to promote the learning tasks in target domain plays a critical role in real-world applications. Recently, lots of deep learning approaches based on autoencoders have achieved significance performance in domain adaptation. However, most existing methods focus on minimizing the distribution divergence by putting the source data and target data together to learn global feature representations, while do not take the local relationship between instances of the same category in different domains into account. To address this problem, we propose a novel Semi-Supervised Representation Learning framework via Dual Autoencoders for domain adaptation, named SSRLDA. More specifically, \textcolor{red}{we extract richer feature representations by learning the global and local feature representations simultaneously using two novel autoencoders}, which are referred to as marginalized denoising autoencoder with adaptation distribution (MDA$_{ad}$) and multi-class marginalized denoising autoencoder (MMDA) respectively. Meanwhile, we \textcolor{red}{adopt an iterative strategy} to make full use of label information to optimize feature representations. Experimental results show that our proposed approach outperforms several state-of-the-art baseline methods.